Anomaly Detection in Cyber-Physical Systems using Machine Learning Techniques

Thesis Abstract

Abstract
Cyber-physical systems (CPS) are ubiquitous in modern society, integrating physical processes with computing and communication elements to enhance efficiency and functionality. However, the interconnected nature of CPS exposes them to various security threats, including anomalies that can disrupt operations and compromise system integrity. Anomaly detection is crucial for safeguarding CPS against such threats, and machine learning techniques have shown promise in effectively identifying and mitigating anomalies in complex systems. This thesis focuses on the application of machine learning methods for anomaly detection in CPS, aiming to enhance system security and reliability. Chapter 1 provides an introduction to the research topic, outlining the background of the study, problem statement, objectives, limitations, scope, significance, structure of the thesis, and key definitions. The literature review in Chapter 2 presents a comprehensive analysis of existing research on anomaly detection in CPS, covering various machine learning algorithms and methodologies employed in this domain. Chapter 3 details the research methodology, including data collection procedures, feature selection techniques, model training and evaluation methods, and performance metrics used to assess the effectiveness of anomaly detection algorithms in CPS. The chapter also discusses the experimental setup and validation strategies employed to validate the proposed approach. In Chapter 4, the findings of the study are extensively discussed, highlighting the performance of machine learning models in detecting anomalies in CPS datasets. The chapter provides insights into the strengths and limitations of different algorithms, as well as the impact of various factors on anomaly detection accuracy and efficiency. The discussion also addresses challenges encountered during the research process and proposes potential solutions for future work in this area. Finally, Chapter 5 presents the conclusions drawn from the study and summarizes the key findings and contributions of the research. The chapter also offers recommendations for further research to enhance anomaly detection capabilities in CPS using machine learning techniques. Overall, this thesis contributes to the growing body of knowledge on cybersecurity in CPS and provides valuable insights for researchers, practitioners, and policymakers seeking to enhance the security and resilience of interconnected systems in the digital age.